CN112776673B - Intelligent network fuel cell automobile real-time energy optimization management system - Google Patents
Intelligent network fuel cell automobile real-time energy optimization management system Download PDFInfo
- Publication number
- CN112776673B CN112776673B CN202011409457.XA CN202011409457A CN112776673B CN 112776673 B CN112776673 B CN 112776673B CN 202011409457 A CN202011409457 A CN 202011409457A CN 112776673 B CN112776673 B CN 112776673B
- Authority
- CN
- China
- Prior art keywords
- time
- soc
- value
- power
- fuel cell
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 239000000446 fuel Substances 0.000 title claims abstract description 201
- 238000005457 optimization Methods 0.000 title claims abstract description 144
- 238000005096 rolling process Methods 0.000 claims abstract description 72
- 238000013461 design Methods 0.000 claims abstract description 15
- 210000004027 cell Anatomy 0.000 claims description 197
- 229910052739 hydrogen Inorganic materials 0.000 claims description 69
- 239000001257 hydrogen Substances 0.000 claims description 69
- UFHFLCQGNIYNRP-UHFFFAOYSA-N Hydrogen Chemical compound [H][H] UFHFLCQGNIYNRP-UHFFFAOYSA-N 0.000 claims description 68
- 238000000034 method Methods 0.000 claims description 35
- 238000005070 sampling Methods 0.000 claims description 29
- 238000004364 calculation method Methods 0.000 claims description 20
- 230000005540 biological transmission Effects 0.000 claims description 9
- 230000007704 transition Effects 0.000 claims description 9
- 230000009471 action Effects 0.000 claims description 6
- 230000005484 gravity Effects 0.000 claims description 6
- 210000002569 neuron Anatomy 0.000 claims description 6
- 238000013528 artificial neural network Methods 0.000 claims description 5
- 230000001133 acceleration Effects 0.000 claims description 4
- 230000004913 activation Effects 0.000 claims description 3
- 238000000638 solvent extraction Methods 0.000 claims description 3
- 238000012546 transfer Methods 0.000 claims description 3
- 230000006855 networking Effects 0.000 abstract description 8
- 238000004422 calculation algorithm Methods 0.000 description 18
- 230000006870 function Effects 0.000 description 18
- 230000008569 process Effects 0.000 description 12
- 230000000694 effects Effects 0.000 description 8
- 238000010586 diagram Methods 0.000 description 6
- 238000009826 distribution Methods 0.000 description 6
- 238000004088 simulation Methods 0.000 description 4
- 230000004044 response Effects 0.000 description 3
- 230000002457 bidirectional effect Effects 0.000 description 2
- 230000008859 change Effects 0.000 description 2
- 238000004891 communication Methods 0.000 description 2
- 238000011161 development Methods 0.000 description 2
- 238000004134 energy conservation Methods 0.000 description 2
- 238000005265 energy consumption Methods 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 238000010438 heat treatment Methods 0.000 description 2
- 230000003993 interaction Effects 0.000 description 2
- 238000011160 research Methods 0.000 description 2
- 238000013473 artificial intelligence Methods 0.000 description 1
- 230000001174 ascending effect Effects 0.000 description 1
- 230000002238 attenuated effect Effects 0.000 description 1
- 230000008901 benefit Effects 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 238000002485 combustion reaction Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 230000005611 electricity Effects 0.000 description 1
- 230000007613 environmental effect Effects 0.000 description 1
- 238000003912 environmental pollution Methods 0.000 description 1
- 150000002431 hydrogen Chemical class 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 239000000463 material Substances 0.000 description 1
- 230000005012 migration Effects 0.000 description 1
- 238000013508 migration Methods 0.000 description 1
- 230000008092 positive effect Effects 0.000 description 1
- 238000002360 preparation method Methods 0.000 description 1
- 230000002035 prolonged effect Effects 0.000 description 1
- 230000001172 regenerating effect Effects 0.000 description 1
- 238000012549 training Methods 0.000 description 1
- 239000002699 waste material Substances 0.000 description 1
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 1
Images
Classifications
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B60—VEHICLES IN GENERAL
- B60L—PROPULSION OF ELECTRICALLY-PROPELLED VEHICLES; SUPPLYING ELECTRIC POWER FOR AUXILIARY EQUIPMENT OF ELECTRICALLY-PROPELLED VEHICLES; ELECTRODYNAMIC BRAKE SYSTEMS FOR VEHICLES IN GENERAL; MAGNETIC SUSPENSION OR LEVITATION FOR VEHICLES; MONITORING OPERATING VARIABLES OF ELECTRICALLY-PROPELLED VEHICLES; ELECTRIC SAFETY DEVICES FOR ELECTRICALLY-PROPELLED VEHICLES
- B60L58/00—Methods or circuit arrangements for monitoring or controlling batteries or fuel cells, specially adapted for electric vehicles
- B60L58/40—Methods or circuit arrangements for monitoring or controlling batteries or fuel cells, specially adapted for electric vehicles for controlling a combination of batteries and fuel cells
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
- G05B13/00—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion
- G05B13/02—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric
- G05B13/04—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric involving the use of models or simulators
- G05B13/042—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric involving the use of models or simulators in which a parameter or coefficient is automatically adjusted to optimise the performance
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02T—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO TRANSPORTATION
- Y02T10/00—Road transport of goods or passengers
- Y02T10/60—Other road transportation technologies with climate change mitigation effect
- Y02T10/70—Energy storage systems for electromobility, e.g. batteries
Landscapes
- Engineering & Computer Science (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Computation (AREA)
- Sustainable Energy (AREA)
- Power Engineering (AREA)
- Transportation (AREA)
- Mechanical Engineering (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Sustainable Development (AREA)
- Medical Informatics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Software Systems (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Automation & Control Theory (AREA)
- Fuel Cell (AREA)
- Electric Propulsion And Braking For Vehicles (AREA)
Abstract
An intelligent network-connected fuel cell automobile real-time energy optimization management system belongs to the field of fuel cell automobile optimization control. The invention aims to provide a layered real-time energy rolling optimization control intelligent networking fuel cell automobile real-time energy optimization management system for a fuel cell automobile. The invention designs a macroscopic long-time-domain average traffic flow velocity trajectory prediction module, designs a microscopic short-time-domain vehicle speed prediction module, establishes a fuel cell vehicle power system model facing energy optimization control, establishes an energy optimization management problem, designs an upper-layer trajectory rolling optimization controller by using long-time-domain preview information, designs a lower-layer energy rolling optimization controller by using short-time-domain preview information, and transmits a solved control input sequence signal to a power execution control unit of a fuel cell vehicle. The invention excavates the energy-saving space of the fuel cell automobile in the intelligent network traffic environment and obviously improves the fuel economy of the fuel cell automobile in the intelligent network environment.
Description
Technical Field
The invention belongs to the field of fuel cell automobile optimization control.
Background
With the increasing automobile holding capacity in China, the problems of energy consumption and environmental pollution become more serious, and energy conservation and environmental protection become the core problems of China's attention. The fuel cell automobile is a novel clean energy automobile which is very environment-friendly. Compared with the traditional internal combustion engine automobile, the working process of the fuel cell is not limited by Carnot cycle, the energy conversion efficiency is high, and the end discharge product is water, so that zero emission and zero pollution can be realized. The large-scale application of hydrogen energy and fuel cell automobiles is realized, and the quantity of the fuel cell automobiles is about 100 thousands; and simultaneously, the key technology of the fuel cell core is completely mastered, and a complete fuel cell material, component and system preparation and production industrial chain is established.
The fuel cell vehicle is powered by two power sources, namely a fuel cell and a power cell. The fuel cell is the core energy of a fuel cell hybrid power system, has higher power density, plays a main role in ensuring the normal running process of a vehicle, but has slower dynamic response speed and is difficult to adapt to the frequent change of a load; the power battery is an auxiliary energy source of a fuel battery hybrid power system, has low power density but high dynamic response speed, can better cope with the situation of large load change, and provides partial energy which cannot be provided by the fuel battery when an automobile runs. Meanwhile, when the automobile is in a braking state, the working mode of the motor of the automobile is changed from a motor mode to a generator mode, and the power battery can recover regenerative braking energy generated during braking of the automobile and charge the power battery for self, so that the waste of energy can be avoided to a certain extent. The two power sources work cooperatively, so that the defects of low response speed of the fuel cell and low power density of the power cell are overcome.
The energy flow direction of the fuel cell automobile is complex, the working efficiency of the fuel cell automobile shows a trend of ascending first and then descending along with the increase of power, the fuel cell automobile is not a simple linear relation, the power of each power source needs to be reasonably distributed in an energy management strategy in a fuel cell automobile system, the fuel cell can work in a high-efficiency interval as far as possible, and the power cell can fully utilize and recover braking energy. The energy management strategy is the core for improving the fuel economy, so that the design of an excellent energy management system is of great importance to fuel cell vehicles. In addition, with the rapid development of artificial intelligence, sensor technology and cloud computing, the intelligent networked automobile is an inevitable trend of future automobile development, and the intelligent networked automobile senses the surrounding environment by using a sensor of the intelligent networked automobile and performs information interaction with other vehicles through communication equipment to obtain multi-source driving preview information. By utilizing the multi-source preview information, the vehicle control system performs energy optimization management on the fuel cell vehicle in a predictive manner by combining the reference information given by the cloud computing, and can provide a larger promotion space for the fuel economy of the vehicle.
The patent CN108944900A discloses a fuel cell vehicle energy management control method, which formulates a power distribution strategy of a fuel cell and a power cell according to the power required by the current vehicle in driving and in combination with the state of charge of the power cell. However, the invention is a rule-based instantaneous energy management strategy, and can only make a decision on energy distribution according to the current vehicle state, and cannot optimize the working efficiency of the fuel cell and the power cell. In addition, the invention does not relate to an automobile energy management strategy in an intelligent networking environment, and the energy-saving potential of the vehicle is not fully exploited.
The patent CN109795373A discloses a fuel cell commercial vehicle energy management control method based on durability, and the method adopts a fuzzy logic strategy to distribute the output power of a fuel cell and a power cell, so that the fuel economy of a fuel cell vehicle is better improved. However, the strategy needs to design corresponding power distribution fuzzy logic under the condition of known fixed driving conditions, can only perform off-line operation, and has great application limitation in the face of unknown complex traffic environments with high uncertainty.
The patent CN110696815A discloses a method for predicting energy management of a grid-connected hybrid electric vehicle, which takes a gasoline-electric hybrid electric vehicle as a research object and distributes energy of an engine and a power battery by introducing grid-connected information. However, the invention does not relate to an energy optimization management strategy for a fuel cell vehicle, the fuel cell is different from an engine, the fuel cell cannot be started and stopped frequently during working, a stricter constraint condition is provided, and the energy management design difficulty of the fuel cell vehicle is higher compared with an oil-electricity hybrid power vehicle.
In summary, although the presently disclosed patents have been directed to some solutions for energy management of fuel cell vehicles, basically, the methods are based on rule design, and it is difficult to obtain the optimal fuel economy, and in order to obtain the optimal fuel economy, the calculation amount of the algorithm to be used is large, and it is necessary to know complete working condition information in advance, and only offline operation is possible. In order to solve the contradiction between the real-time performance and the energy-saving effect of the optimization algorithm and further excavate the energy-saving potential of the fuel cell vehicle in the intelligent network traffic environment, the challenge is also filled with the design of the real-time energy optimization management system of the intelligent network fuel cell vehicle.
Disclosure of Invention
The invention aims to predict vehicle speed information in a short time domain by using a historical vehicle speed sequence, and provides a layered real-time energy rolling optimization control intelligent network fuel cell vehicle real-time energy optimization management system by combining vehicle speed preview information in a long time domain and a multi-power source system model.
The method comprises the following steps:
the method comprises the following steps: designing a macroscopic long time domain average traffic flow velocity trajectory prediction module;
the method is characterized in that:
step two: vehicle speed prediction module for designing microscopic short time domain
The neural network comprises a three-layer structure, namely an input layer, a hidden layer and an output layer. The input vector is defined as m (k), and the output vector is defined as
The structure of the BPNN can be represented by a discrete model with weights and thresholds
Wherein,w1is the weight between the input layer and the hidden layer, w2Is the weight between the hidden layer and the output layer, b1Is a threshold for hidden layer neurons, b2Is a threshold for output layer neurons, m (k) represents the input sequence of historical vehicle speeds,
a sequence of predicted vehicle speeds representing an output, g (h) being a hidden layer to output layer activation function, transfer function thereof
Step three: establishing fuel cell automobile power system model oriented to energy optimization control
3.1 building a longitudinal driving dynamics model of the automobile
Fuel cell vehicle parameters: motor transmission efficiency etat_veh(%), mass coefficient σ of rotating elementveh(-) -acceleration of gravity g (m/s)2) Air resistance coefficient CD_veh(-) -air density ρair(kg/m3) Mass m of automobileveh(kg), frontal area Aveh(m2) Sliding resistance coefficient f (-) and road surface gradient thetaroad(-)；
Power demand of vehicle
Wherein P isveh_reqIs the required power of the vehicle, f is the coefficient of sliding resistance, ηt_vehIs the motor transmission efficiency, σvehIs the mass coefficient of the rotating element, mvehIs the mass of the vehicle, g is the acceleration of gravity, θroadIs the road surface gradient, AvehIs the frontal area, rho, of the automobileairIs the air density, CD_vehIs the coefficient of air resistance and is,
is the speed V of the vehiclevehDifferentiation with respect to time t;
3.2 establishing a fuel cell stack efficiency model
Hydrogen consumption of fuel cell
Wherein, Pfc_reqIs the output power of the fuel cell, etafc_stIs the efficiency of the operation of the fuel cell,
is the lower heating value of hydrogen;
3.3 establishing a power battery SOC model
Pbatt_req＝Pveh_req-Pfc_req (5)
Wherein, Pbatt_reqIs the output power of the power cell;
the SOC dynamic equation of the power battery is
Wherein, Voc_battIs the open circuit voltage, R, of the power cellint_battIs the internal resistance, Q, of the power cellbattIs the total charge of the power cell,
is the derivative of the power cell state of charge, SOC;
step four: establishing an energy optimization management problem
4.1 output power P of the Fuel cellfc_reqThe state equation is:
the above formula can be represented by
Represents;
minimizing the prediction time domain [ t ]0,tf]Hydrogen consumption of the system:
wherein J is the total hydrogen consumption in the prediction time domain under the condition of satisfying the system terminal constraint, t0Is the starting time, t, of the prediction time domainfIs the end time of the prediction time domain, U is the control input variable, U is the value set of the control input variable,
representing the hydrogen consumption of the system at time t as a function of a variable related to the control input u (t) at time t, where u is Pfc_reqThe state variable x is equal to SOC, phi (x (t)f) Terminal constraints that are state variables;
4.2 the following constraints are satisfied:
(1) the output power constraint of the fuel cell needs to be satisfied:
Pfc_low≤Pfc_req(t)≤Pfc_up (10)
wherein, Pfc_lowIs the minimum output power, P, of the fuel cellfc_upIs the maximum output power, P, of the fuel cellfc_req(t) is the output power of the fuel cell at time t;
(2) the dynamic equation and state constraint of the power battery SOC need to be satisfied:
therein, SOCbeginIs a dynamic electricitySOC value of the pool at initial time, SOClowIs the minimum value of SOC of the power batteryupIs the maximum value of the power battery SOC, SOC (t) represents the value of the power battery SOC at the time t, and SOC (t)0) Representing the value of the power battery SOC at the initial moment, SOC (t)f) Representing the value of the SOC of the power battery at the terminal moment;
(3) need to satisfy power constraints of power cells
Pbatt_low≤Pbatt_req(t)≤Pbatt_up (12)
Wherein, Pbatt_lowIs the maximum charging power, P, of the power batterybatt_upIs the maximum discharge power, P, of the power batterybatt_req(t) is the output power of the power battery at the time t;
(4) the required power of the automobile during operation needs to be satisfied
Pveh_req(t)＝Pbatt_req(t)+Pfc_req(t) (13)
Wherein, Pveh_req(t) is the power demand of the vehicle at time t, Pbatt_req(t) is the output power of the power battery at the time t;
step five: design upper layer SOC track rolling optimization controller by using long time domain preview information
5.1 upper SOC trajectory rolling optimization controller optimization problem
Predicting the time domain [ t ] at the time scale0,m,tf,m]Is dispersed into NmEqual portion, wherein, t0,mIs the starting time, t, of the prediction time domainf,mThe end time of the prediction time domain is recorded as k e {1,2m+1} to obtain an optimization objective:
wherein J is the total hydrogen consumption of the system at all sampling moments under the condition of satisfying the terminal constraint, phi (x (N)m+1)) is a terminal constraint for the state variable,
the representative hydrogen consumption is a function related to a control input u (k) at the moment k, delta t is a sampling time interval between two adjacent vehicle speed information, and a control variable u (k) is the output power P of the fuel cell at the moment kfc_req_m(k)；
The concrete constraint conditions met are as follows:
(1) satisfying the output power constraint of the fuel cell:
Pfc_low≤Pfc_req_m(k)≤Pfc_up (15)
(2) the dynamic equation and state constraint of the power battery SOC in the time domain are met:
therein, SOCm(k +1) is the power battery SOC at time km(k) The SOC value of the power battery at the next moment in the time domain, namely the SOC value of the power battery at the moment k +1, V is obtained after the control input actionoc_batt_m(k) Is the open-circuit voltage, R, of the power battery at the time kint_batt_m(k) Is the internal resistance, P, of the power battery at the time kveh_req_m(k) Is the power demand, P, of the vehicle at time kfc_req_m(k) Is the output power, SOC, of the fuel cell at the time km(k) Is the state of charge (SOC) of the power battery at the time km(1) Is the value of the initial time of the power battery SOC in the time domain, SOCm(Nm+1) is the value of the power battery SOC terminal time in the time domain;
(3) the output power constraint of the power battery is met:
Pbatt_low≤Pbatt_req_m(k)≤Pbatt_up (17)
wherein, Pbatt_req_m(k) Is the output power of the power battery at the moment k in the time domain;
(4) meet the power demand of the running of the automobile
Pveh_req_m(k)＝Pfc_req_m(k)+Pbatt_req_m(k) (18)；
5.2 partitioning the grid with respect to System State and control variables
Dividing the state variable power battery into 81 state grids; the output power of the fuel cell is increased by 81 grids of control variables from the beginning;
5.3 calculate cost
Under the action of a control variable u (k), a state variable x (k) can obtain a new state variable x (k +1) after being calculated by a state transition equation, from 1 moment, different control variable grids act on the state variable grids to obtain a state variable grid at the next moment, corresponding cost J (k) is generated, meanwhile, the new control variable grid acts on the state variable grid at the moment, the cost J (k +1) corresponding to the next moment is generated until the whole driving cycle working condition is calculated, and the generated cost can be calculated by a formula
Calculating, and storing cost generated by each front-to-back iterative calculation in a grid;
5.4 determining optimal decisions
Determining a terminal time k-NmThe value of the state variable of +1, i.e. x (N)m+1) corresponding to the initial objective function J (N)mWhen +1) ═ 0, from the time immediately before the end time, there are:
wherein, J*(k) Represents the minimum value of the hydrogen consumption when the system state variable is x (k) at the k-th time. L (x (k), u (k)) represents the hydrogen consumption generated by the system at the state variable x (k) under the action of the control input u (k) at the k moment, namely the state transition cost, J*(k +1) is the minimum value of hydrogen consumption when the system state variable is x (k +1) at the last moment, and the corresponding state variable which enables the minimum value of the cost function is selected from each moment, so that the optimal state variable sequence { x } can be obtained*(1),x*(2),...,x*(k) H, the optimal power battery SOC sequence SOC*；
Step six: design of lower-layer energy rolling optimization controller by using short-time-domain preview information
6.1 receiving SOC in prediction time Domain at Current sampling time*A track sequence, reading the SOC value of the current power battery;
6.2 lower-layer energy rolling optimization controller optimization problem
Micro short time domain t0,n,tf,n]The vehicle speed preview information is dispersed into NnEqual parts, with the discrete time s e {1,2n+1}, where t is0,nIs the starting time, t, of the prediction horizonf,nIs the end time of the prediction time domain, and an optimization objective function is obtained:
wherein u isdIs the control input, U, in the time domaindIs a control input value set in the value range of the control input in the time domain, and I is the SOC and SOC of the power battery at all the sampling moments of the system under the condition of meeting the constraint of the system terminal*Sum of squares of differences and sum of hydrogen consumption, SOCn(s) is the value of the power battery SOC at s time in the time domain, phi (x)d(Nn+1)) is the terminal constraint of the state variable, ud(s) is the value of the control input in the time domain at time s,
representative of hydrogen consumption is the control input ud(s) function of interest, control variable selection ud＝Pfc_req_n(s) state variable selection of xd＝SOCn(s), the specific constraints satisfied are:
(1) satisfying the output power constraint of the fuel cell:
Pfc_low＜Pfc_req_n(s)＜Pfc_up (21)
(2) the dynamic equation and the state constraint of the power battery SOC are met:
therein, SOCn(s +1) is the SOC of the power battery at s moment in the time domainn(s) the value of the power battery SOC at the next moment in the time domain, namely the value of the power battery SOC at the moment of s +1, V, obtained after the control input actionoc_batt_n(s) is the open circuit voltage of the power cell at time s in the time domain, Pveh_req_n(s) is the power demand of the vehicle at time s in the time domain, Pfc_req_n(s) is the output power of the fuel cell at time s in the time domain, Rint_batt_n(s) is the internal resistance, SOC, of the power battery at s time in the time domainn(1) Is the value of the power battery SOC at the initial moment in the time domain, SOCn(Nn+1) is the value of the power battery SOC at the terminal time in the time domain;
(3) the output power constraint of the power battery is met:
Pbatt_low≤Pbatt_req_n(s)≤Pbatt_up (23)
(4) the required power during the operation of the automobile is met:
Pveh_req_n(s)＝Pfc_req_n(s)+Pbatt_req_n(s) (24)
6.3 constructing Hamiltonian
H(xd(s),ud(s),λ(s),s)＝(SOCn(s)-SOC*(s))2·Δt
+((WH2_fc(ud(s))·Δt+λ(s)ΔSOCn(s), (25)
Wherein, H (x)d(s),ud(s), λ(s), s) represent the values x of the Hamiltonian and the state variables at the time sd(s) control input value u at time sd(s), the value λ(s) of the covariate at time s is related to the current time s, and the necessary conditions that the optimization needs to meet are as follows:
λ(s+1)＝λ(s)+Δλ(s)·Δt, (26)
wherein, Delta lambda(s) is the difference value of the covariates at two adjacent time points,
and the value representing the partial derivative of the Hamiltonian at the s moment on the SOC of the power battery is obtained, wherein lambda (s +1) is the value of the covariate at the next moment in the time domain obtained after the covariate lambda(s) at the s moment is calculated, namely the value of the covariate at the s +1 moment. At the same time, optimal control input
The hamiltonian needs to be guaranteed to be minimum at each sampling moment, that is, the following formula needs to be satisfied:
wherein,
is the optimal value of the state variable at time s,
is the optimum value of the control input at time s, λ*(s) is the optimum value of the covariate at time s,
representing the optimal Hamiltonian and the optimal value of the state variable at s moment
Optimal value of control input at time s
Optimum lambda of the covariate at s-time*(s) is related to the current time instant s,
representing the optimum values of the Hamiltonian and the state variables at time s
Control input value u at time sd(s) optimal value λ of the covariate at s moment*(s) is related to the current time instant s;
6.4 solving optimal control input sequences
(1) State variable SOC for setting initial time0And calculating initial value lambda of the covariate at the initial time by bisection method0；
(2) The control input when the Hamilton function value is minimum is the optimal control input
Ud(s)＝[ud_low(s):Δud(s):ud_up(s)], (28)
Wherein, Δ ud(s) is the difference of two adjacent equal parts of the control input at the time s in the time domain, ud_up(s) is the maximum value of the control input within the constraint range at time s, ud_low(s) is the minimum value of the control input within the constraint range, Ud(s) is a value set of the control input at time s in the time domain;
(3) according to optimal control input action
Calculating the SOC value and the lambda value of the co-modal variable at the next sampling moment according to the result of the state transition equation, and repeating the step 2) until the last sampling moment;
(4) judging the final tail end boundary error value of the SOC value, and ending if the error is in the set rangeCalculation otherwise requiring re-input of λ0Determining the value of the covariance variable lambda within the error allowable range by a dichotomy within the value range set by lambda, repeating the step 2), and obtaining an optimal control input sequence after all the calculations are completed;
step seven: and transmitting the solved control input sequence signal to a power execution control unit of the fuel cell automobile.
The invention provides an intelligent network-connected fuel cell automobile real-time energy optimization management system facing a fuel cell automobile in an intelligent network-connected environment, solves the contradiction between optimization instantaneity and energy-saving effect in the existing energy optimization research algorithm, further excavates the energy-saving space of the fuel cell automobile in the intelligent network-connected traffic environment, and obviously improves the fuel economy of the fuel cell automobile in the intelligent network-connected environment.
Drawings
FIG. 1 is a schematic diagram of a power and transmission part of a fuel cell vehicle;
FIG. 2 is a schematic diagram of an intelligent networked fuel cell vehicle real-time energy optimization management system;
FIG. 3 is a flow chart of the operation of the real-time energy optimization management system of the intelligent networked fuel cell vehicle;
FIG. 4 is a schematic illustration of the BPNN algorithm predicting short term vehicle speeds;
FIG. 5 is a schematic view of a multi-time scale vehicle speed information prediction;
FIG. 6 is a graph of operating efficiency of a fuel cell;
FIG. 7 is a graph of the relationship between the internal resistance and the SOC of the power battery;
FIG. 8 is a vehicle speed profile of a selected congestion operating condition driving cycle (LA 92);
FIG. 9 shows the optimal SOC of the power battery calculated by the upper-layer SOC track rolling optimization controller under the driving cycle (LA92) under the congestion condition*A trajectory;
FIG. 10 is a plot of vehicle speed versus a BPNN predicted vehicle speed for a congestion operating condition driving cycle (LA 92);
FIG. 11 is a reference optimal S calculated by the upper SOC trajectory rolling optimization controller under a driving cycle under congestion conditions (LA92)OC*A comparison graph of the track and an actual SOC curve calculated by a lower-layer energy rolling optimization controller;
fig. 12 is a graph of results of fuel cell output power and power cell output power calculated by the lower energy rolling optimization controller, and required power of the vehicle under a congestion condition driving cycle (LA 92);
fig. 13 is a calculation time curve required by the energy rolling optimization controller to solve the output power of the fuel cell and the power cell at each sampling time in the driving cycle under the congestion condition (LA 92);
fig. 14 is a hydrogen consumption curve calculated by the intelligent networked fuel cell vehicle real-time energy optimization management system under a congestion condition driving cycle (LA 92);
fig. 15 is a comparison graph of the hydrogen consumption calculated by the intelligent networked fuel cell vehicle real-time energy optimization management system under the congestion condition driving cycle (LA92), and the hydrogen consumption calculated by the rule-based energy management strategy and the result of the offline global optimal hydrogen consumption.
Detailed Description
The main challenges and problems of the existing fuel cell automobile under the intelligent network traffic environment include: 1. macroscopic and microscopic traffic aiming information obtained in the internet environment is dynamically updated along with the migration of time and space, and a challenge is how to design an efficient energy optimization management system by utilizing the abundant multi-source aiming information to realize the prediction energy conservation of a fuel cell automobile; 2. the performance of the fuel cell is attenuated due to frequent starting and stopping of the fuel cell, the State of Charge (SOC) of the power cell has higher working efficiency only within a proper working range, the starting and stopping of the fuel cell stack and the SOC range of the power cell need to be strictly limited in the driving process of the fuel cell hybrid electric vehicle, and the design of a multi-power-source real-time energy optimization algorithm with constraint conditions is a challenge.
The method comprises the following steps:
the method comprises the following steps: and designing a macroscopic long time domain average traffic flow velocity trajectory prediction module. The network terminal generates average traffic flow velocity track preview information of a macroscopic long time domain (600 seconds) by collecting an electronic map, a Global Positioning System (GPS) and driving information of other vehicles in internet traffic by using cloud computing resources, and dynamically updates the average traffic flow velocity track preview information in real time according to the frequency of once every 300 seconds.
Step two: and designing a vehicle speed prediction module in a microscopic short time domain. And generating the speed information of the pre-aiming microscopic short time domain (5 seconds) by combining the average traffic flow velocity track information of the macroscopic long time domain and utilizing a Back Propagation Neural Network (BPNN) according to the historical speed sequence of the vehicle in the past 5 seconds, and dynamically updating in real time according to the frequency of once every 1 second.
Step three: and establishing an energy optimization control-oriented fuel cell automobile power system model. Establishing a longitudinal driving dynamics model of the automobile, a fuel cell stack efficiency model and a power cell SOC model; and calculating the required power of the automobile according to the average traffic flow velocity track information of the macroscopic long time domain and the speed preview information of the microscopic short time domain.
Step four: and establishing energy optimization management problem description. Selecting control input variables, establishing energy optimization management problem description, and determining constraint conditions of optimization problems.
Step five: and designing an upper-layer SOC track rolling optimization controller by using the long time domain preview information. According to the average traffic flow velocity track information of the macroscopic long time domain provided by the cloud computing resources and the required power calculated in the second step, the rolling optimization problem on-line solving algorithm based on a Dynamic Programming (DP) algorithm is provided with the aim of minimizing the hydrogen consumption of the fuel cell vehicle under the condition of ensuring that the system terminal constraint is met, and the optimal power cell SOC track in the time domain is solved by utilizing the cloud computing resources.
Step six: and designing a lower-layer energy rolling optimization controller by using short time domain preview information. Power battery reference SOC track (SOC) optimized by SOC track rolling optimization controller*) The SOC provided by the rolling optimization controller of the upper-layer SOC track by the actual power battery SOC in the microscopic short time domain while the system terminal constraint is met is used as the reference input of the rolling optimization controller of the lower-layer energy*The minimum tracking error and hydrogen consumption are taken as targets, in PontUnder the theoretical framework of the Riyagin Maximum Principle (PMP), an online solving algorithm for the rolling optimization problem of the fuel cell automobile energy is provided, and the output power sequences of the fuel cell and the power cell in the short time domain, namely the optimal control input sequence of the system, are solved.
Step seven: and transmitting the solved control input sequence signal to a power execution control unit of the fuel cell automobile.
Step eight: and carrying out experimental simulation, and evaluating the energy-saving effect of the designed real-time energy optimization management system.
The information collected by the invention comprises position information and gradient information collected by a GPS, route information collected by an electronic map, intersection traffic light information collected by internet communication, driving state information of other vehicles and the like.
The invention provides a real-time energy optimization management system of an intelligent network-connected fuel cell automobile and aims at the fuel cell automobile in the intelligent network-connected environment, the traffic preview information provided by cloud computing resources is utilized, the problem of performance attenuation caused by the dynamic characteristics and frequent start and stop of the fuel cell is considered, and the hydrogen-saving potential of the fuel cell automobile in the intelligent network-connected traffic environment is exploited.
The multi-scale dynamic traffic preview internet connection information related by the invention comprises the following steps: average traffic flow velocity track information in a macroscopic long time domain (minute level) and vehicle speed preview information in a microscopic short time domain (second level). The long-time-domain average traffic flow velocity trajectory information is processed by the cloud computing data processing center, provided to the fuel cell automobile, and dynamically updated in a fixed time (usually 5-10 minutes). The short-time-domain vehicle speed track preview information (usually 5-10 seconds) can be obtained by real-time learning through a machine learning method (the invention takes BPNN as an example) by combining data such as a historical vehicle speed sequence of a target vehicle and long-time-domain average traffic flow velocity track information.
The invention combines traffic information of two different time scales of a macroscopic long time domain and a microscopic short time domain, aims at the multi-power source energy coordination optimization problem of a fuel cell automobile, considers the system safety constraint under the congestion driving state, and provides a hierarchical real-time energy rolling optimization control method of the fuel cell automobile, wherein the hierarchical real-time energy rolling optimization control method comprises 1) an upper-layer SOC track rolling optimization controller and 2) a lower-layer energy rolling optimization controller, and the description is as follows:
(1) the upper layer SOC track rolling optimization controller: by utilizing cloud computing resources, taking the average traffic flow velocity track information of the front 600 seconds updated every 300 seconds as the reference input of the system, aiming at the optimal fuel economy under the condition of meeting the system terminal constraint, and solving the SOC track SOC of the power battery under the condition of minimum hydrogen consumption by applying a DP algorithm*。
(2) The lower energy rolling optimization controller: and the vehicle-mounted controller is combined with the average traffic flow velocity track information of the upper layer, the traffic light information of the intersection, the gradient information of the current road and the historical vehicle speed sequence of the past 5 seconds, and a BPNN algorithm is utilized to obtain a microscopic short-time-domain vehicle speed sequence which predicts the future 5 seconds at the frequency of once every 1 second. Combining the short-term vehicle speed prediction sequence obtained by the vehicle-mounted controller, rolling the upper layer SOC track to optimize the optimal power battery SOC track SOC obtained by the controller*As the reference input of the layer, the SOC given by the rolling optimization controller of the power battery SOC on the upper layer SOC track in the microscopic short time domain is used for ensuring that the system terminal constraint is met*The trace tracking error and the hydrogen consumption are minimum as targets, and the output power of the fuel cell and the power cell is solved by utilizing a PMP algorithm in the short time domain to obtain an optimal power distribution scheme.
The structure diagram of the power and transmission part of the fuel cell automobile is shown in figure 1, the fuel cell and the power cell are respectively connected to a circuit bus through a unidirectional DC/DC converter and a bidirectional DC/DC converter, the circuit bus is connected with a motor through the bidirectional DC/AC converter, and the motor drives wheels to rotate so as to provide power for the running of the automobile.
The invention discloses a simplified diagram of an intelligent networked fuel cell automobile real-time energy optimization management system, which is shown in figure 2. The specific implementation mode is as follows: the network terminal collects information such as position information, destination information and interaction information of other networked vehicles of the automobile and uploads the information to the cloud computing data processing center. After the cloud computing data processing center processes the acquired information, the information is processed once every 300 secondsThe new frequency generates average traffic flow velocity track information (600 seconds) of a macroscopic long time domain and sends the average traffic flow velocity track information to a target vehicle, and the vehicle-mounted controller is combined with the average traffic flow velocity track information of the upper layer, traffic light information of an intersection, gradient information of a current road and a historical vehicle speed sequence of the past 5 seconds, and predicts a microscopic short time domain vehicle speed sequence of the future 5 seconds according to the frequency updated every 1 second by utilizing a BPNN algorithm. And then establishing an automobile longitudinal driving dynamics model, a fuel cell stack efficiency model, a battery SOC model and a real-time energy optimization management system, establishing problem description of a real-time energy optimization management problem, and determining constraint conditions of the optimization problem. In a real-time energy optimization management system, an upper-layer SOC track rolling optimization controller combines average traffic flow speed information and required power of a macroscopic long time domain of 600 seconds in front, which is updated once every 300 seconds, and under the condition of meeting system terminal constraints, the aim of minimum hydrogen consumption is fulfilled, cloud computing resources solve the SOC track SOC of a power battery in the time domain by adopting a DP algorithm*. The lower layer energy rolling optimization controller optimizes the optimal SOC track (SOC)*) As a reference input, the vehicle-mounted controller ensures that the system terminal constraint is met and simultaneously uses the SOC given by the power battery SOC to the upper-layer SOC track rolling optimization controller in a microscopic short time domain*The trace tracking error and the hydrogen consumption are minimum as targets, and a fuel cell power and power cell power sequence of the system, namely an optimal control input sequence, is obtained. And finally, transmitting the obtained optimal control input sequence signal to a power execution control unit of the fuel cell automobile. And carrying out experimental simulation on the designed system, and verifying the energy-saving effect of the designed system on the fuel cell automobile. Specifically, the method comprises the following steps:
the work flow chart of the real-time energy optimization management system of the intelligent networked fuel cell automobile is shown in fig. 3, and the method specifically comprises the following steps:
1. and designing a macroscopic long time domain average traffic flow velocity trajectory prediction module.
The network terminal generates average traffic flow velocity track information of a preview macro long time domain (600 seconds) by collecting driving information of an electronic map, a GPS and other vehicles in internet traffic by using cloud computing resources, transmits the average traffic flow velocity track information to the fuel cell automobile, and dynamically updates the average traffic flow velocity track information in real time according to the frequency of once every 300 seconds.
2. Vehicle speed prediction module for designing microscopic short time domain
The vehicle-mounted controller combines the average traffic flow velocity track information of a macroscopic long time domain, the traffic light information of a crossing, the gradient information of the current road and the historical vehicle speed sequence of the past 5 seconds, generates the vehicle speed preview information of a microscopic short time domain (5 seconds) by using a BPNN algorithm, and dynamically updates in real time according to the frequency of once every 1 second.
The schematic diagram of the BPNN algorithm for predicting the short-term vehicle speed is shown in FIG. 4, and the neural network comprises three layers, namely an input layer, a hidden layer and an output layer. The input vector is defined as m (k), and the output vector is defined as
The structure of the BPNN may be represented by a discrete model with weights and thresholds, as shown in equation (1):
wherein, w1Is the weight between the input layer and the hidden layer, w2Is the weight between the hidden layer and the output layer, b1Is a threshold for hidden layer neurons, b2Is a threshold for output layer neurons, m (k) represents the input sequence of historical vehicle speeds,
representing the output sequence of predicted vehicle speeds, g (h) is a hidden layer to output layer activation function.
The transfer function is shown in formula (2):
and training through various typical working condition data to obtain a neural network vehicle speed prediction model. In the model, a vehicle historical vehicle speed sequence of 5 seconds is input according to the frequency of once every 1 second, namely, a vehicle speed sequence of 5 seconds in the future can be predicted in a rolling mode. Fig. 10 is a comparison graph of a vehicle speed curve of a congestion condition driving cycle (LA92) and a vehicle speed curve predicted by BPNN, and it can be seen from the graph that the vehicle speed curve predicted by the BPNN algorithm can be matched with the vehicle speed curve of an actual driving cycle to a high degree, and a good prediction effect is achieved.
3. And establishing an energy optimization control-oriented fuel cell automobile power system model.
Establishing a longitudinal driving dynamics model, a fuel cell stack efficiency model and a dynamic electricity of the automobile
And the pool SOC model is used for calculating the required power of the automobile by combining the average traffic flow speed track information of the macroscopic long time domain and the speed preview information of the microscopic short time domain.
3.1 building a longitudinal driving dynamics model of the automobile
And calculating the required power of the automobile during running according to the speed preview information. The fuel cell vehicle parameters of this patent are shown in table 1:
table 1: fuel cell vehicle parameter table
| Meaning of variables | Symbol and unit |
| Efficiency of motor transmission | ηt_veh(％) |
| Mass coefficient of rotating element | σveh(-) |
| Acceleration of gravity | g(m/s2) |
| Coefficient of air resistance | CD_veh(-) |
| Density of air | ρair(kg/m3) |
| Automobile mass | mveh(kg) |
| Frontal area | Aveh(m2) |
| Coefficient of sliding resistance | f(-) |
| Road surface gradient | θroad(-) |
And (3) calculating the required power of the vehicle by using a formula (3) according to the average traffic flow velocity track information of the macroscopic long time domain provided by the cloud computing resources.
Wherein P isveh_reqIs the required power of the vehicle, f is the coefficient of sliding resistance, ηt_vehIs the motor transmission efficiency, σvehIs the mass coefficient of the rotating element, mvehIs the mass of the vehicle, g is the acceleration of gravity, θroadIs the road surface gradient, AvehIs the frontal area, rho, of the automobileairIs the density of air, CD_vehIs the coefficient of air resistance and is,
is the speed V of the vehiclevehDifferentiation with respect to time t.
3.2 establishing a fuel cell stack efficiency model
The fuel cell is the main energy source for driving the fuel cell vehicle to run, and the operating efficiency curve of the fuel cell related to the invention obtained by an experimental calibration mode is shown in fig. 6. Hydrogen consumption of fuel cell
The calculation formula is as follows:
wherein, Pfc_reqIs the output power, eta, of the fuel cellfc_stIs the efficiency of the operation of the fuel cell,
is the lower heating value of hydrogen, with a value of 120000J/g, which indicates that burning 1g of hydrogen will produce 120000J of energy.
3.3 establishing a power battery SOC model
The power battery is an auxiliary energy source of the fuel cell vehicle, and the output power of the power battery is the difference between the required power of the vehicle and the output power of the fuel cell, as shown in formula 5:
Pbatt_req＝Pveh_req-Pfc_req, (5)
wherein, Pbatt_reqIs the output power of the power battery.
The SOC dynamic equation of the power battery is
Wherein, Voc_battIs the open circuit voltage, R, of the power cellint_battIs the internal resistance, Q, of the power cellbattIs the total charge of the power cell,
is the derivative of the power cell state of charge SOC.
4. Establishing energy-optimized management problem descriptions
Selecting control input variables, establishing energy optimization management problem description, and determining constraint conditions of optimization problems.
4.1 building energy-optimized management problem descriptions
The relation curve graph of the internal resistance and the SOC of the power battery is shown in FIG. 7, when the SOC state is too high or too low, the internal resistance is large, and the working efficiency is low; the life and operating performance of fuel cells can be degraded by frequent start-stops during vehicle operation. In combination with the characteristics of the power source, the start and stop of the fuel cell stack and the SOC range of the power cell need to be strictly limited during the optimization process. The invention selects the SOC of the power battery as the state variable, and can obtain the open-circuit voltage V by analyzing the working principle of the fuel battery hybrid power systemoc_battInternal resistance R of power batteryint_battAre all functions related to the SOC of the power battery, thus a plurality of control variables in the dynamic equation can be simplified into one control variable, namely the output power P of the fuel batteryfc_reqThe state equation is as follows:
the above formula can be represented by
And (4) showing.
The optimization target is to minimize the prediction time domain [ t ] under the condition of satisfying the system terminal constraint0,tf]Hydrogen consumption of the system:
wherein J is the total hydrogen consumption in the prediction time domain under the condition of satisfying the system terminal constraint, t0Is the starting time, t, of the prediction time domainfIs predictive of the time domainThe termination time, U is the control input variable, U is the value set of the control input variable,
representing the hydrogen consumption of the system at time t as a function of a variable related to the control input u (t) at time t, where u is Pfc_reqThe state variable x is equal to SOC, phi (x (t)f) Is a terminal constraint of the state variable.
4.2 determining constraints for optimization problems
The intelligent networked fuel cell automobile real-time energy optimization management system needs to meet the following constraint conditions:
(1) the output power constraint of the fuel cell needs to be satisfied:
Pfc_low≤Pfc_req(t)≤Pfc_up, (9)
wherein, Pfc_lowIs the minimum output power, P, of the fuel cellfc_upIs the maximum output power, P, of the fuel cellfc_req(t) is the output power of the fuel cell at time t.
The dynamic equation and state constraint of the power battery SOC need to be satisfied:
therein, SOCbeginIs the SOC value of the power battery at the initial time, SOClowIs the minimum value of SOC of the power batteryupIs the maximum value of the power battery SOC, SOC (t) represents the value of the power battery SOC at the time t, and SOC (t)0) Representing the value of the power battery SOC at the initial moment, SOC (t)f) Representing the value of the SOC of the power battery at the terminal moment;
4.3 need to meet Power constraints of Power batteries
Pbatt_low≤Pbatt_req(t)≤Pbatt_up, (11)
Wherein, Pbatt_lowIs the maximum charging power, P, of the power batterybatt_upIs the maximum discharge power, P, of the power batterybatt_req(t) is the output power of the power battery at the time t;
4.4 Power requirement for vehicle operation
Pveh_req(t)＝Pbatt_req(t)+Pfc_req(t)， (12)
Wherein, Pveh_req(t) is the power demand of the vehicle at time t, Pbatt_reqAnd (t) is the output power of the power battery at the time t.
And solving the optimal control quantity of the system by combining the determined optimization problem and the constraint condition. A multi-time scale vehicle speed information prediction diagram is shown in FIG. 5. On the upper layer, the cloud computing resource predicts the average traffic flow velocity track information of a macroscopic long time domain 600 seconds ahead according to the frequency of once every 300 seconds, and transmits the predicted traffic flow velocity track information to an upper SOC track rolling optimization controller and a lower vehicle-mounted controller; and in the lower layer, the vehicle-mounted controller is combined with the traffic flow velocity track information, the traffic light information of the intersection, the gradient information of the current road and the historical vehicle speed sequence of the past 5 seconds, predicts the vehicle speed preview information of the microscopic short time domain 5 seconds ahead according to the frequency of once every 1 second, and transmits the vehicle speed preview information of the short time domain to the lower layer energy rolling optimization controller. And designing an upper-layer SOC track rolling optimization controller (step 5) and a lower-layer energy rolling optimization controller (step 6) according to the vehicle speed preview information of the upper layer and the lower layer with different time scales. The specific description is as follows:
5. and designing an upper-layer SOC track rolling optimization controller by using the long time domain preview information.
Combining the average traffic flow velocity trajectory information and the required power of the macroscopic long time domain, and aiming at the optimal fuel economy, designing an upper-layer SOC trajectory rolling optimization controller to solve the optimal SOC trajectory SOC of the power battery in the time domain*. The method specifically comprises the following steps:
5.1 Upper layer SOC trajectory roll optimization controller optimization problem description
The upper-layer SOC track rolling optimization controller utilizes the average traffic flow speed track information of the macroscopic long time domain of the front 600 seconds, which is provided by the cloud computing resources and is updated every 300 secondsAnd under the condition of ensuring that the constraint of the system terminal is met, the minimum hydrogen consumption is taken as a target, and the cloud computing resources are utilized to carry out optimization solution. In the solving process, data needs to be sampled, and the predicted time domain [ t ] under the time scale is used0,m,tf,m]Is dispersed into NmEqual parts, wherein, t0,mIs the starting time, t, of the prediction time domainf,mThe end time of the prediction time domain is recorded as k e {1,2m+1}, obtaining an optimization target:
wherein J is the total hydrogen consumption of the system at all sampling moments under the condition of satisfying the terminal constraint, phi (x (N)m+1)) is a terminal constraint for the state variable,
the representative hydrogen consumption is a function related to a control input u (k) at the moment k, delta t is a sampling time interval between two adjacent vehicle speed information, and a control variable u (k) is the output power P of the fuel cell at the moment kfc_req_m(k) In the track rolling optimization process, it is required to ensure that the initial SOC value of the power battery is equal to the terminal SOC value, because when the initial SOC of the power battery is different from the final SOC, equivalent hydrogen consumption is generated. If the final SOC value is higher than the initial SOC value, a part of energy of the fuel cell is considered to be stored in the power battery and not really consumed, and similarly, when the final SOC value of the power battery is lower than the initial SOC value, the power battery is considered to additionally consume energy instead of the fuel cell. In order to avoid the interference of the power battery with additional equivalent hydrogen consumption on the hydrogen consumption result and facilitate the comparison of the hydrogen consumption result, the final SOC needs to be constrained to be the same as the initial SOC value. The specific constraints that need to be satisfied are:
1. the output power constraint of the fuel cell needs to be satisfied:
Pfc_low≤Pfc_req_m(k)≤Pfc_up； (14)
2. the dynamic equation and state constraint of the power battery SOC in the time domain need to be satisfied:
therein, SOCm(k +1) is the power battery SOC at time km(k) The SOC value of the power battery at the next moment in the time domain, namely the SOC value of the power battery at the moment k +1, V is obtained after the control input actionoc_batt_m(k) Is the open-circuit voltage, R, of the power battery at the time k in the time domainint_batt_m(k) Is the internal resistance, P, of the power battery at the time kveh_req_m(k) Is the power demand, P, of the vehicle at time kfc_req_m(k) Is the output power, SOC, of the fuel cell at time k in this time domainm(k) Is the state of charge (SOC) of the power battery at the k moment in the time domainm(1) Is the value of the initial time of the power battery SOC in the time domain, SOCm(Nm+1) is the value of the power battery SOC terminal time in the time domain;
3. the output power constraint of the power battery needs to be satisfied:
Pbatt_low≤Pbatt_req_m(k)≤Pbatt_up, (16)
wherein, Pbatt_req_m(k) Is the output power of the power battery at the time k in the time domain;
4. the required power of the automobile during operation needs to be satisfied
Pveh_req_m(k)＝Pfc_req_m(k)+Pbatt_req_m(k)， (17)
5.2 partitioning the grid with respect to System states and control variables
To apply the DP method, the SOC of the state variable power cell is incremented from 0.3 to 0.7 with an increase of 0.005 per cell, dividing out 81 state grids; the fuel cell output power was increased from 5kW with an increase of 0.5625kW per division to 81 grids of control variables of 50 kW.
5.3 calculating cost
Under the action of the control variables u (k), the state variables x (k) are calculated by the state transition equationThen, a new state variable x (k +1) is obtained. Starting from time 1, the different control variable grids acting on the state variable grid will obtain the state variable grid at the next time, resulting in the corresponding cost j (k). And simultaneously, a new control variable grid is acted on the state variable grid at the moment, and the cost J (k +1) corresponding to the next moment is generated until the working condition of the whole driving cycle is calculated. The generated cost can be formulated
And calculating, and storing cost generated by each forward-backward iterative calculation in the grid.
5.4 determining optimal decisions
Determining the optimal control input sequence needs to be accomplished by iterative computation from back to front. The optimization problem aims to minimize the hydrogen consumption of the fuel cell vehicle under the condition of meeting terminal constraints. First, a terminal time k is determined to be Nm+1 value of the state variable (taking the value of SOC)begin) I.e. x (N)m+1) corresponding to the initial objective function J (N)mWhen +1) ═ 0, from the time immediately before the end time, there are:
wherein, J*(k) Represents the minimum value of the hydrogen consumption when the system state variable is x (k) at the k-th time. L (x (k), u (k)) represents the hydrogen consumption generated by the system under the action of the state variable x (k) through the control input u (k) at the k-th moment, namely the state transition cost. J is a unit of*And (k +1) is the minimum value of the hydrogen consumption when the system state variable is x (k +1) at the last moment. Selecting the corresponding state variable which enables the cost function to be minimum at each moment, and obtaining the optimal state variable sequence { x }*(1),x*(2),...,x*(k) H, the optimal power battery SOC sequence SOC*。
6. And designing a lower-layer energy rolling optimization controller by using short time domain preview information.
Rolling optimizing control of upper layer SOC trackOptimal SOC trajectory SOC given by a system*As reference input, vehicle speed preview information of a front 5-second micro short time domain updated every 1 second obtained by BPNN algorithm is combined, and the actual power battery SOC in the micro short time domain is used while the system terminal constraint is metnSOC given by rolling optimization controller for upper SOC track*And designing a lower-layer energy rolling optimization controller by taking the minimum track tracking error and hydrogen consumption as targets, and solving an output power sequence of the fuel cell and the power cell in the time domain by using a vehicle-mounted controller, namely an optimal control input sequence of the system. The method specifically comprises the following steps:
6.1 receiving SOC in prediction time Domain at Current sampling time*And a track sequence for reading the SOC value of the current power battery.
6.2 underlying energy roll optimization controller optimization problem description
The lower-layer energy rolling optimization controller ensures that the actual power battery SOC in a microscopic short time domain meets the system terminal constraintnSOC given by rolling optimization controller for upper SOC track*And solving the output power of the fuel battery and the power battery by taking the minimum trace tracking error and the minimum hydrogen consumption as targets. In the solving process, a microscopic short time domain t needs to be obtained0,n,tf,n]The vehicle speed preview information is dispersed into NnEqual parts, with the discrete time s e {1,2n+1}, where t is0,nIs the starting time, t, of the prediction horizonf,nIs the end time of the prediction time domain, and an optimization objective function is obtained:
wherein u isdIs the control input, U, in the time domaindIs a control input value set in the value range of the control input in the time domain, and I is the SOC and SOC of the power battery at all sampling moments of the system under the condition of meeting the constraint of a system terminal*Sum of squares of difference and sum of hydrogen consumption, SOCn(s) is the value of the SOC of the power battery at s time in the time domain, [ phi ] (xd(Nn+1)) is the terminal constraint of the state variable, ud(s) is the value of the control input in the time domain at time s,
representative of hydrogen consumption is the control input ud(s) function of interest, control variable chosen ud＝Pfc_req_n(s) state variable selection of xd＝SOCn(s)。
The specific constraints that need to be satisfied are:
(1) the output power constraint of the fuel cell needs to be satisfied:
Pfc_low＜Pfc_req_n(s)＜Pfc_up； (20)
(2) the dynamic equation and state constraint of the power battery SOC need to be satisfied:
therein, SOCn(s +1) is the SOC of the power battery at s moment in the time domainn(s) the value of the power battery SOC at the next moment in the time domain, namely the value of the power battery SOC at the moment of s +1, V, obtained after the control input actionoc_batt_n(s) is the open circuit voltage of the power cell at time s in the time domain, Pveh_req_n(s) is the power demand of the vehicle at time s in the time domain, Pfc_req_n(s) is the output power of the fuel cell at time s in the time domain, Rint_batt_n(s) is the internal resistance, SOC, of the power battery at s time in the time domainn(1) Is the value of the power battery SOC at the initial moment in the time domain, SOCn(Nn+1) is the value of the power battery SOC at the terminal moment in the time domain, and the energy rolling optimization controller also needs to meet the condition that the SOC value of the system terminal is equal to the initial SOC value so as to avoid the interference generated by equivalent hydrogen consumption;
3. the output power constraint of the power battery needs to be satisfied:
Pbatt_low≤Pbatt_req_n(s)≤Pbatt_up； (22)
4. the required power when the automobile runs needs to be met:
Pveh_req_n(s)＝Pfc_req_n(s)+Pbatt_req_n(s). (23)
6.3 constructing Hamiltonian
The hamiltonian is constructed as follows:
wherein, H (x)d(s),ud(s), λ(s), s) represent the values x of the Hamiltonian and the state variables at the time sd(s) control input of value u at time sd(s), the value of the covariate at time s, λ(s), is related to the current time s.
The requirements to be met by optimization are as follows:
λ(s+1)＝λ(s)+Δλ(s)·Δt,
wherein, Delta lambda(s) is the difference value of the covariates at two adjacent time points,
and the value representing the partial derivative of the Hamiltonian at the s moment on the SOC of the power battery is obtained, wherein lambda (s +1) is the value of the covariate at the next moment in the time domain obtained after the covariate lambda(s) at the s moment is calculated, namely the value of the covariate at the s +1 moment.
At the same time, optimal control input
The hamiltonian needs to be guaranteed to be minimum at each sampling moment, that is, the following formula needs to be satisfied:
wherein,
is the optimal value of the state variable at time s,
is the optimum value of the control input at time s, λ*(s) is the optimum value of the covariate at time s,
representing the optimal Hamiltonian and the optimal value of the state variable at s moment
Optimal value of control input at time s
Optimum lambda of the covariate at s-time*(s) is related to the current time instant s,
representing the optimum values of the Hamiltonian and the state variables at time s
Controlling the value u of the input at time sd(s) optimal value λ of the covariate at s moment*(s) is related to the current time s, and equations (24) - (26) are actually a boundary value problem between two fixed endpoints, which is solved by satisfying the requirements of PMP to obtain the optimal control input sequence.
6.4 solving optimal control input sequences
(1) State variable SOC for setting initial time0And calculating initial value lambda of the covariate at the initial time by dichotomy0。
(2) Within the allowable range of the control input, the control input is divided into 100 equal parts, the Hamiltonian is calculated by acting on each sampling moment, and the control input with the smallest Hamiltonian value is the optimal control input
Ud(s)＝[ud_low(s):Δud(s):ud_up(s)],
Wherein, Δ ud(s) is the difference of two adjacent equal parts of the control input at the time s in the time domain, ud_up(s) is the maximum value of the control input within the constraint range at time s, ud_low(s) is the minimum value of the control input within the constraint range, UdAnd(s) is a value set of the control input in the time domain at the time of s.
(3) According to the optimal control input action
Under the calculation of the results of the state transition equation
Repeating the step 2) until the last sampling moment according to the SOC value and the lambda value of the covariate at the sampling moment;
(4) judging the final tail end boundary error value of the SOC value, finishing the calculation if the error is in a set range, otherwise, re-inputting the lambda0And determining the value of the covariate lambda within the error allowable range by a dichotomy within the value range set by the lambda. And (5) repeating the step 2), and obtaining the optimal control input sequence after all the calculations are completed.
7. And transmitting the solved optimal control input sequence signal to a power execution control unit of the fuel cell automobile, transmitting the optimal control input sequence signal obtained by the lower-layer energy rolling optimization controller to the power execution control unit of the fuel cell automobile through a data bus, and acting on a fuel cell hybrid power system to enable the fuel cell automobile to distribute power according to the result of algorithm calculation, thereby finally improving the fuel economy of the fuel cell automobile.
8. And performing experimental simulation, evaluating the energy-saving effect of the designed real-time energy optimization management system, and selecting a vehicle speed curve of a congestion driving cycle working condition (LA92) to verify the effectiveness of the designed system. The LA92 driving cycle curve duration 1435 seconds contains abundant working condition information such as low speed, frequent start-stop.
According to the simulation result, the real-time energy optimization management system for the intelligent networked fuel cell automobile, provided by the invention, has the following advantages:
(1) upper layer of designed real-time energy optimization management system of intelligent network-connected fuel cell automobile
The SOC track rolling optimization controller and the lower-layer energy rolling optimization controller fully utilize networking information of a macroscopic long time domain and vehicle speed preview information of a microscopic short time domain to obtain a power battery SOC track corresponding to the optimal hydrogen consumption in a macroscopic prediction time domain, provide sufficient reference information for the lower-layer energy rolling optimization controller, and further develop the energy-saving potential of the fuel cell vehicle.
FIG. 8 is a vehicle speed profile for a selected congestion operating cycle (LA 92). FIG. 11 is a reference optimal SOC calculated by the upper SOC trajectory rolling optimization controller under a driving cycle under congestion conditions (LA92)*The comparison graph of the track and the actual SOC curve calculated by the lower-layer energy rolling optimization controller shows that the tracking effect on the SOC is good, and the SOC value is always in the constraint range. Fig. 14 is a hydrogen consumption curve calculated by the real-time energy optimization management system of the intelligent networked fuel cell vehicle under the congestion condition driving cycle (LA92), and fig. 15 is a comparison graph of the hydrogen consumption calculated by the real-time energy optimization management system of the intelligent networked fuel cell vehicle under the congestion condition driving cycle (LA92), the hydrogen consumption calculated by the energy management strategy based on the rules, and the results of the offline global optimal hydrogen consumption. It can be seen from the figure that the hydrogen consumption calculated by the real-time energy optimization management strategy of the fuel cell vehicle connected with the internet designed by the invention is 169.93g, the hydrogen consumption calculated by the rule-based energy management strategy is 199.32g, the theoretical optimal hydrogen consumption obtained by the offline global optimal strategy is 164.52g, and the comparison shows that the energy consumption of the designed system is saved by 17.3% compared with the energy management strategy based on the ruleCompared with the energy management strategy based on the rules, the hydrogen quantity and the fuel economy are obviously improved, the difference between the theoretical optimal hydrogen consumption calculated by the offline global optimal strategy and the theoretical optimal hydrogen consumption is only 3.2 percent, the theoretical optimal hydrogen consumption is close to the theoretical optimal hydrogen consumption, the designed system is very effective, abundant networking information can be fully utilized, the hydrogen-saving potential of the fuel cell automobile is excavated, and the fuel economy of the automobile is greatly improved.
(2) The reasonable and effective distribution of energy of the automobile in the running process is ensured, the power of the fuel cell is always more than 5kW, and the condition that the fuel cell stack is stopped in the running process of the automobile is avoided. Meanwhile, the SOC of the power battery is always in the restraint range in the running process of the vehicle, so that the power battery is prevented from working in a low-efficiency interval, and the service lives of the fuel battery and the power battery are prolonged.
FIG. 12 is a graph showing the results of the fuel cell output power and the power cell output power calculated by the lower energy rolling optimization controller for the required power of the vehicle in the driving cycle under the congested operating conditions (LA92), and FIG. 9 is a graph showing the optimal SOC of the power cell calculated by the upper SOC trajectory rolling optimization controller in the driving cycle under the congested operating conditions (LA92)*The track is shown in FIG. 11, which is a reference optimal SOC calculated by an upper-layer SOC track rolling optimization controller under a congestion condition driving cycle (LA92)*The comparison graph of the actual SOC curve calculated by the track and the lower-layer energy rolling optimization controller shows that the tracking effect of the SOC is good, and the value of the SOC is always kept between 0.45 and 0.52, namely the optimal efficiency interval of the power battery, which shows that the energy distribution among the fuel battery, the power battery and the vehicle transmission system is reasonable and effective, and the cruising ability and the service life of the fuel battery vehicle are further improved.
(3) The designed lower-layer energy rolling optimization controller is high in calculation speed, and real-time performance of system solution is guaranteed. The longest calculation time for solving the power battery SOC track at all sampling moments of the upper-layer SOC track rolling optimization controller is 41.63 seconds which are far shorter than 300 seconds, and the updating frequency of the upper-layer SOC track rolling optimization controller for once in 300 seconds can be met; fig. 13 is a calculation time curve required by the energy rolling optimization controller to solve the output power of the fuel cell and the power cell at each sampling time under the driving cycle (LA92) under the congestion condition, and it can be seen from the graph that the longest calculation time of all sampling times of the lower-layer energy rolling optimization controller is 0.052 seconds, which is much shorter than 1 second, and can satisfy the update frequency of sampling once every 1 second, thereby ensuring the real-time performance and the rapidity of system solution.
The invention provides a hierarchical real-time energy rolling optimization control method for the fuel cell vehicle by combining multi-scale networking prediction information, so that the calculation burden of the real-time energy management optimization of the fuel cell vehicle is effectively reduced, and the energy efficiency of the fuel cell vehicle is greatly improved.
The invention has the positive effects that:
1. an intelligent networking fuel cell automobile real-time energy optimization management system and a system working process are designed, and by utilizing multi-scale networking information, compared with a rule-based energy management strategy, the fuel economy is improved by 17.3%, and the energy-saving space of the fuel cell automobile is further excavated;
2. aiming at the problem of predicting energy-saving optimization of a multi-power source fuel cell vehicle under multi-scale networking information, a layered real-time energy rolling optimization control method for the fuel cell vehicle is provided, and the obtained fuel economy is close to the theoretical optimal economy; meanwhile, the maximum solving time for solving the output power of the fuel cell and the power cell at all sampling moments of the lower layer is only 0.052 second, which is far less than 1 second, and the updating frequency of sampling once every 1 second can be met. The method not only effectively utilizes the preview information of different time scales, fully excavates the hydrogen-saving potential of the fuel cell automobile, but also can meet the calculation requirement of real-time performance.
3. By means of macroscopic long-time-domain internet preview information, an SOC track rolling optimization controller is designed on the upper layer, a power battery SOC track corresponding to optimal fuel economy in the time domain is solved by means of cloud computing resources, track information is sent to the lower energy rolling optimization controller to provide abundant reference information for the power battery SOC track, and the energy-saving space of the fuel battery automobile is fully improved.
4. And designing an energy rolling optimization controller on the lower layer by using microscopic short-time-domain preview information and combining SOC track reference information obtained by rolling optimization of the upper layer, and rapidly and efficiently solving the output power of the fuel cell and the power cell by using a vehicle-mounted controller.
Claims (1)
1. An intelligent network fuel cell automobile real-time energy optimization management system,
the method comprises the following steps: designing a macroscopic long time domain average traffic flow velocity trajectory prediction module;
the method is characterized in that:
step two: vehicle speed prediction module for designing microscopic short time domain
The neural network comprises a three-layer structure, namely an input layer, a hidden layer and an output layer; the input vector is defined as m (k), and the output vector is defined as
The structure of the BPNN can be represented by a discrete model with weights and thresholds
Wherein, w1Is the weight between the input layer and the hidden layer, w2Is the weight between the hidden layer and the output layer, b1Is a threshold for hidden layer neurons, b2Is a threshold for output layer neurons, m (k) represents the input sequence of historical vehicle speeds,
a sequence of predicted vehicle speeds representing an output, g (h) being a hidden layer to output layer activation function, transfer function thereof
Step three: establishing fuel cell automobile power system model oriented to energy optimization control
3.1 building a longitudinal driving dynamics model of the automobile
Fuel cell vehicle parameters: motor transmission efficiency etat_veh(%), mass coefficient σ of rotating elementveh(-) -acceleration of gravity g (m/s)2) Air resistance coefficient CD_veh(-) -air density ρair(kg/m3) Mass m of automobileveh(kg) frontal area Aveh(m2) Sliding resistance coefficient f (-) and road surface gradient thetaroad(-)；
Power demand of vehicle
Wherein P isveh_reqIs the required power of the vehicle, f is the coefficient of sliding resistance, ηt_vehIs the motor transmission efficiency, σvehIs the mass coefficient of the rotating element, mvehIs the mass of the vehicle, g is the acceleration of gravity, θroadIs the road surface gradient, AvehIs the frontal area, rho, of the automobileairIs the air density, CD_vehIs the coefficient of air resistance and is,
is the speed V of the vehiclevehDifferentiation with respect to time t;
3.2 establishing a fuel cell stack efficiency model
Hydrogen consumption of fuel cell
Wherein, Pfc_reqIs the output power of the fuel cell, etafc_stIt is the efficiency of operation of the fuel cell,
is low in hydrogenA calorific value;
3.3 establishing a power battery SOC model
Pbatt_req＝Pveh_req-Pfc_req (5)
Wherein, Pbatt_reqIs the output power of the power cell;
the SOC dynamic equation of the power battery is
Wherein, Voc_battIs the open circuit voltage, R, of the power cellint_battIs the internal resistance, Q, of the power cellbattIs the total charge of the power cell,
is the derivative of the power cell state of charge, SOC;
step four: establishing an energy optimization management problem
4.1 output power P of the Fuel cellfc_reqThe state equation is:
the above formula can be represented by
Represents;
minimizing the prediction time domain [ t ]0,tf]Hydrogen consumption of the system:
wherein J is the total hydrogen consumption in the prediction time domain under the condition of satisfying the system terminal constraint, t0Is the starting time, t, of the prediction time domainfIs predicting the termination of the time domainIn between, U is the control input variable, U is the value set of the control input variable,
representing the hydrogen consumption of the system at time t as a function of a variable related to the control input u (t) at time t, where u is Pfc_reqThe state variable x is equal to SOC, phi (x (t)f) Terminal constraints that are state variables;
4.2 the following constraints are satisfied:
(1) the output power constraint of the fuel cell needs to be satisfied:
Pfc_low≤Pfc_req(t)≤Pfc_up (10)
wherein, Pfc_lowIs the minimum output power, P, of the fuel cellfc_upIs the maximum output power, P, of the fuel cellfc_req(t) is the output power of the fuel cell at time t;
(2) the dynamic equation and state constraint of the power battery SOC need to be satisfied:
therein, SOCbeginIs the SOC value of the power battery at the initial time, SOClowIs the minimum value of SOC of the power batteryupIs the maximum value of the power battery SOC, SOC (t) represents the value of the power battery SOC at the time t, and SOC (t)0) Representing the value of the power battery SOC at the initial moment, SOC (t)f) Representing the value of the SOC of the power battery at the terminal moment;
(3) need to satisfy power constraints of power cells
Pbatt_low≤Pbatt_req(t)≤Pbatt_up (12)
Wherein, Pbatt_lowIs the maximum charging power, P, of the power batterybatt_upIs the maximum discharge power, P, of the power batterybatt_req(t) is the output power of the power battery at the time t;
(4) the required power of the automobile during operation needs to be satisfied
Pveh_req(t)＝Pbatt_req(t)+Pfc_req(t) (13)
Wherein, Pveh_req(t) is the power demand of the vehicle at time t, Pbatt_req(t) is the output power of the power battery at the time t;
step five: design upper layer SOC track rolling optimization controller by using long time domain preview information
5.1 upper SOC trajectory rolling optimization controller optimization problem
Predicting the time domain [ t ] at the time scale0,m,tf,m]Is dispersed into NmEqual parts, wherein, t0,mIs the starting time, t, of the prediction time domainf,mThe end time of the prediction time domain is the discrete time k epsilon {1,2m+1}, obtaining an optimization target:
wherein J is the total hydrogen consumption of the system at all sampling moments under the condition of satisfying the terminal constraint, phi (x (N)m+1)) is a terminal constraint for the state variable,
the representative hydrogen consumption is a function related to a control input u (k) at the moment k, delta t is a sampling time interval between two adjacent vehicle speed information, and a control variable u (k) is the output power P of the fuel cell at the moment kfc_req_m(k)；
The specific constraint conditions met are:
(1) satisfying the output power constraint of the fuel cell:
Pfc_low≤Pfc_req_m(k)≤Pfc_up (15)
(2) the dynamic equation and state constraint of the power battery SOC in the time domain are met:
therein, SOCm(k +1) is the power battery SOC at time km(k) The SOC value of the power battery at the next moment in the time domain, namely the SOC value of the power battery at the moment k +1, V is obtained after the control input actionoc_batt_m(k) Is the open-circuit voltage, R, of the power battery at the time kint_batt_m(k) Is the internal resistance, P, of the power battery at the time kveh_req_m(k) Is the power demand, P, of the vehicle at time kfc_req_m(k) Is the output power, SOC, of the fuel cell at time k in this time domainm(k) Is the state of charge (SOC) of the power battery at the k moment in the time domainm(1) Is the value of the initial time of the power battery SOC in the time domain, SOCm(Nm+1) is the value of the power battery SOC terminal time in the time domain;
(3) the output power constraint of the power battery is satisfied:
Pbatt_low≤Pbatt_req_m(k)≤Pbatt_up (17)
wherein, Pbatt_req_m(k) Is the output power of the power battery at the moment k in the time domain;
(4) meet the power demand of the running of the automobile
Pveh_req_m(k)＝Pfc_req_m(k)+Pbatt_req_m(k) (18)；
5.2 partitioning the grid with respect to System states and control variables
Dividing the state variable power battery into 81 state grids; the fuel cell output power is increased by 81 grids of control variables from the beginning;
5.3 calculate cost
Under the action of the control variable u (k), the state variable x (k) can obtain a new state variable x (k +1) after the calculation of the state transition equation, from the moment 1, different control variable grids act on the state variable grids to obtain the state variable grids at the next moment, the corresponding cost J (k) is generated, and simultaneously, the new control variable grids act on the state variable grids at the moment to generate the generation corresponding to the next momentThe price cost J (k +1) is calculated till the working condition of the whole driving cycle is completed, and the generated cost can be expressed by a formula
Calculating, and storing cost generated by each time of forward-backward iterative calculation in a grid;
5.4 determining optimal decisions
Determining a terminal time k-NmThe value of the state variable of +1, i.e. x (N)m+1) corresponding to the initial objective function J (N)mWhen +1) ═ 0, from the time immediately before the end time, there are:
wherein, J*(k) Represents the minimum value of the hydrogen consumption when the system state variable is x (k) at the k-th moment; l (x (k), u (k)) represents the hydrogen consumption generated by the system at the k-th moment under the action of the state variable x (k) through the control input u (k), namely the state transition cost, J*(k +1) is the minimum value of hydrogen consumption when the system state variable is x (k +1) at the last moment, and the corresponding state variable which enables the minimum value of the cost function is selected from each moment, so that the optimal state variable sequence { x } can be obtained*(1),x*(2),...,x*(k) I.e. the optimal power battery SOC sequence SOC*；
Step six: design of lower-layer energy rolling optimization controller by using short-time-domain preview information
6.1 receiving SOC in prediction time Domain at Current sampling time*A track sequence, reading the SOC value of the current power battery;
6.2 lower-layer energy rolling optimization controller optimization problem
Micro short time domain t0,n,tf,n]The vehicle speed preview information is dispersed into NnEqual parts, with the discrete time s e {1,2n+1} where t is0,nIs the starting time, t, of the prediction time domainf,nIs the end time of the predicted time domain to obtain an optimized objective function：
Wherein u isdIs the control input, U, in the time domaindIs a control input value set in the value range of the control input in the time domain, and I is the SOC and SOC of the power battery at all sampling moments of the system under the condition of meeting the constraint of a system terminal*Sum of squares of differences and sum of hydrogen consumption, SOCn(s) is the value of the power battery SOC at s time in the time domain, phi (x)d(Nn+1)) is the terminal constraint of the state variable, ud(s) is the value of the control input in the time domain at time s,
representative of hydrogen consumption is the control input ud(s) function of interest, control variable selection ud＝Pfc_req_n(s) state variable selection of xd＝SOCn(s), the specific constraints satisfied are:
(1) satisfying the output power constraint of the fuel cell:
Pfc_low＜Pfc_req_n(s)＜Pfc_up (21)
(2) the dynamic equation and the state constraint of the power battery SOC are met:
therein, SOCn(s +1) is the SOC of the power battery at s moment in the time domainn(s) the SOC value of the power battery at the next moment in the time domain, namely the SOC value of the power battery at the s +1 moment, V, obtained after the control input actionoc_batt_n(s) is the open-circuit voltage of the power cell at s time in the time domain, Pveh_req_n(s) is the power demand of the vehicle at time s in the time domain, Pfc_req_n(s) is the output power of the fuel cell at time s in the time domain, Rint_batt_n(s) is theInternal resistance, SOC, of power battery at time s in time domainn(1) Is the value of the power battery SOC at the initial moment in the time domain, SOCn(Nn+1) is the value of the power battery SOC at the terminal time in the time domain;
(3) the output power constraint of the power battery is met:
Pbatt_low≤Pbatt_req_n(s)≤Pbatt_up (23)
(4) the required power during the operation of the automobile is met:
Pveh_req_n(s)＝Pfc_req_n(s)+Pbatt_req_n(s) (24)
6.3 constructing Hamiltonian
Wherein, H (x)d(s),ud(s), λ(s), s) represent the values x of the Hamiltonian and the state variables at the time sd(s) control input value u at time sd(s), the value λ(s) of the covariate at time s is related to the current time s, and the necessary conditions that the optimization needs to meet are as follows:
λ(s+1)＝λ(s)+Δλ(s)·Δt, (26)
wherein, Delta lambda(s) is the difference value of the covariates at two adjacent time points,
representing the value of partial derivative of the power battery SOC calculated by the Hamiltonian function at the s moment, wherein lambda (s +1) is the value of the covariate at the next moment in the time domain obtained after the covariate lambda(s) at the s moment is calculated, namely the value of the covariate at the s +1 moment; at the same time, optimal control input
The hamiltonian needs to be guaranteed to be minimum at each sampling moment, that is, the following formula needs to be satisfied:
wherein,
is the optimal value of the state variable at time s,
is the optimum value of the control input at time s, λ*(s) is the optimum value of the covariate at time s,
representing the optimal Hamiltonian and the optimal value of the state variable at s moment
Optimal value of control input at time s
Optimum lambda of the covariate at s time*(s) is related to the current time instant s,
representing the optimum values of the Hamiltonian and the state variables at time s
Controlling the value u of the input at time sd(s) optimal value λ of the covariate at s moment*(s) is related to the current time instant s;
6.4 solving optimal control input sequences
(1) Setting state variable SOC at initial time0And calculating the initial time by bisectionInitial value of the covariate0；
(2) The control input when the Hamilton function value is minimum is the optimal control input
Ud(s)＝[ud_low(s):Δud(s):ud_up(s)], (28)
Wherein, Δ ud(s) is the difference of two adjacent equal parts of the control input at the time s in the time domain, ud_up(s) is the maximum value of the control input within the constraint range at time s, ud_low(s) is the minimum value of the control input within the constraint range, Ud(s) is a value set of the control input at time s in the time domain;
(3) according to optimal control input action
Calculating the SOC value and the lambda value of the co-modal variable at the next sampling moment according to the result of the state transition equation, and repeating the step 2) until the last sampling moment;
(4) judging the final tail end boundary error value of the SOC value, finishing the calculation if the error is in a set range, otherwise, re-inputting the lambda0Determining the value of the covariance variable lambda within the error allowable range by a dichotomy within the value range set by lambda, repeating the step 2), and obtaining an optimal control input sequence after all the calculations are completed;
step seven: and transmitting the solved control input sequence signal to a power execution control unit of the fuel cell automobile.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011409457.XA CN112776673B (en) | 2020-12-06 | 2020-12-06 | Intelligent network fuel cell automobile real-time energy optimization management system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011409457.XA CN112776673B (en) | 2020-12-06 | 2020-12-06 | Intelligent network fuel cell automobile real-time energy optimization management system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN112776673A CN112776673A (en) | 2021-05-11 |
| CN112776673B true CN112776673B (en) | 2022-06-21 |
Family
ID=75750741
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011409457.XA Active CN112776673B (en) | 2020-12-06 | 2020-12-06 | Intelligent network fuel cell automobile real-time energy optimization management system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112776673B (en) |
Families Citing this family (18)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113085666B (en) * | 2021-05-18 | 2022-05-10 | 北京理工大学 | Energy-saving driving method for layered fuel cell automobile |
| CN113071506B (en) * | 2021-05-20 | 2022-04-05 | 吉林大学 | Fuel cell automobile energy consumption optimization system considering cabin temperature |
| CN113022384B (en) * | 2021-05-26 | 2021-08-24 | 北京理工大学 | Fuel cell automobile energy management method based on convex optimization |
| CN113263960B (en) * | 2021-06-28 | 2022-08-19 | 太原理工大学 | Self-adaptive energy management method for hydrogen fuel cell automobile |
| CN113246805B (en) * | 2021-07-02 | 2022-07-19 | 吉林大学 | Fuel cell power management control method considering temperature of automobile cockpit |
| CN113561793B (en) * | 2021-08-02 | 2023-09-08 | 昆明理工大学 | Dynamic constraint intelligent fuel cell automobile energy management strategy |
| CN113581019B (en) * | 2021-08-11 | 2023-04-11 | 佛山仙湖实验室 | Energy management method for hybrid drive unmanned mining truck |
| CN113779703B (en) * | 2021-09-22 | 2023-12-22 | 北京理工大学 | Combined optimization method for fuel cell automobile speed planning and energy management |
| CN113815437B (en) * | 2021-10-28 | 2023-11-10 | 吉林大学 | Predictive energy management method for fuel cell hybrid electric vehicle |
| CN114122466B (en) * | 2021-11-11 | 2024-05-31 | 广东汉合汽车有限公司 | Method, system, equipment and medium for optimizing management of fuel cell vehicle battery |
| CN114103971B (en) * | 2021-11-23 | 2023-10-10 | 北京理工大学 | Energy-saving driving optimization method and device for fuel cell automobile |
| CN114154107B (en) * | 2021-12-09 | 2024-07-05 | 重庆金康赛力斯新能源汽车设计院有限公司 | Average energy consumption calculation method and device |
| CN114333364B (en) * | 2021-12-17 | 2022-11-18 | 宁波大学 | Ecological driving method for automatically-driven vehicle to pass through signalized intersection |
| CN114714987B (en) * | 2022-03-08 | 2024-06-25 | 佛山仙湖实验室 | Energy management optimization method and system for fuel cell hybrid electric vehicle |
| CN115230485B (en) * | 2022-07-22 | 2024-08-09 | 大连理工大学 | Fuel cell bus energy management method based on short-term power smooth prediction |
| CN115416503B (en) * | 2022-09-19 | 2024-08-27 | 吉林大学 | Intelligent networking-based energy management method for fuel cell hybrid electric vehicle |
| DE102023112649A1 (en) | 2023-05-12 | 2024-11-14 | Daimler Truck AG | Method for operating an electric drive system of a motor vehicle |
| CN117078030B (en) * | 2023-07-12 | 2024-05-03 | 贵州大学 | Fuel cell bus energy management method based on vehicle speed prediction |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107901776A (en) * | 2017-11-15 | 2018-04-13 | 吉林大学 | Electric automobile composite power source fuel cell hybrid energy system power dividing method |
| CN111891110A (en) * | 2020-05-15 | 2020-11-06 | 吉林大学 | Intelligent network-connected hybrid electric vehicle energy-heat integrated real-time management system |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP6621489B2 (en) * | 2018-02-16 | 2019-12-18 | 本田技研工業株式会社 | Fuel cell system |
-
2020
- 2020-12-06 CN CN202011409457.XA patent/CN112776673B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107901776A (en) * | 2017-11-15 | 2018-04-13 | 吉林大学 | Electric automobile composite power source fuel cell hybrid energy system power dividing method |
| CN111891110A (en) * | 2020-05-15 | 2020-11-06 | 吉林大学 | Intelligent network-connected hybrid electric vehicle energy-heat integrated real-time management system |
Non-Patent Citations (3)
| Title |
|---|
| Dynamic Programming Algorithm for Energy Management Strategy of the Fuel Cell Vehicle;Shengyan Hou等;《2019 3rd Conference on Vehicle Control and Intelligence (CVCI)》;20200109;全文 * |
| Energy Management Strategy of Fuel Cell Hybrid Electric Vehicle Based on Dynamic Programming;Yang hui-ce等;《2020 Chinese Automation Congress (CAC)》;20210129;第3134-3139页 * |
| 混合动力汽车模型预测能量管理研究现状与展望;张风奇等;《机械工程学报》;20190531;第55卷(第010期);第86-108页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN112776673A (en) | 2021-05-11 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112776673B (en) | Intelligent network fuel cell automobile real-time energy optimization management system | |
| Zand et al. | Using adaptive fuzzy logic for intelligent energy management in hybrid vehicles | |
| CN108819934B (en) | Power distribution control method of hybrid vehicle | |
| Liu et al. | Rule-corrected energy management strategy for hybrid electric vehicles based on operation-mode prediction | |
| CN107688343B (en) | Energy control method of hybrid power vehicle | |
| Zhang et al. | A novel learning-based model predictive control strategy for plug-in hybrid electric vehicle | |
| CN113554337B (en) | Plug-in hybrid electric vehicle energy management strategy construction method integrating traffic information | |
| CN114771293A (en) | Fuel cell automobile energy management method based on equivalent consumption minimum strategy | |
| CN115416503A (en) | Fuel cell hybrid electric vehicle energy management method based on intelligent network connection | |
| CN113135113B (en) | Global SOC (System on chip) planning method and device | |
| CN113779703B (en) | Combined optimization method for fuel cell automobile speed planning and energy management | |
| CN112757922A (en) | Hybrid power energy management method and system for vehicle fuel cell | |
| Yang et al. | Energy management of the power-split hybrid electric city bus based on the stochastic model predictive control | |
| CN113815437A (en) | Predictive energy management method for fuel cell hybrid electric vehicle | |
| CN117644783A (en) | Fuel cell automobile energy management method combining working condition prediction and reinforcement learning | |
| Zhang et al. | A Novel Model Predictive Control Based Co-Optimization Strategy for Velocity Planning and Energy Management of Intelligent PHEVs | |
| CN115805840A (en) | Energy consumption control method and system for range-extending type electric loader | |
| Janulin et al. | Energy minimization in city electric vehicle using optimized multi-speed transmission | |
| Liu et al. | Computationally efficient energy management for a parallel hybrid electric vehicle using adaptive dynamic programming | |
| Xia et al. | A predictive energy management strategy for multi-mode plug-in hybrid electric vehicle based on long short-term memory neural network | |
| Yang et al. | A real time multi-objective optimization Guided-MPC strategy for power-split hybrid electric bus based on velocity prediction | |
| CN114291067B (en) | Hybrid electric vehicle convex optimization energy control method and system based on prediction | |
| CN113276829B (en) | Vehicle running energy-saving optimization weight-changing method based on working condition prediction | |
| CN113922375A (en) | Energy management system for wind-light-fire-hydrogen storage new energy system | |
| Zhang et al. | An optimal vehicle speed planning algorithm for regenerative braking at traffic lights intersections based on reinforcement learning |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |